feat(providers): catalog refresh + image_gen.model + OpenRouter caching (WS-6)

Surfaces the v0.13 provider catalog work in Scarf v2.8.0. Five new model IDs
(deepseek/deepseek-v4-pro, x-ai/grok-4.3, openrouter/owl-alpha,
tencent/hy3-preview, arcee/trinity-large-thinking) flow through
models_dev_cache.json on next refresh — no manual catalog entries
needed; the picker reaches them automatically. The grok-4.20-beta →
grok-4.20 rename is handled via a new ModelCatalogService.modelAliases
map plus resolveModelAlias() helper, called from validateModel(),
model(_:_:), and provider(for:) at read time. Lossless: stored configs
are never rewritten.

Vercel AI Gateway is demoted to the bottom of the picker via a new
demotedProviders set + sort-comparator axis (between subscription-gated
and alphabetical). Always-on, no capability gate — sort-order
consistency across Hermes versions.

image_gen.model (top-level v0.13 YAML key) and
openrouter.response_cache.enabled (provisional key shape per
TODO(WS-6-Q1)) are surfaced as new SettingsSection rows in
AuxiliaryTab, capability-gated on hasImageGenModel +
hasOpenRouterResponseCache so pre-v0.13 hosts hide them. Image-gen
picker has a curated 7-entry allowlist (HermesImageGenModel) plus
free-form Custom model ID entry.

CLAUDE.md gains two schema-drift bullets next to the existing
overlayOnlyProviders requirement (modelAliases + demotedProviders
mirror with hermes_cli/providers.py).

Tests: 4 new M0cServicesTests (sort axis, alias resolution + cross-
provider isolation, image-gen allowlist, demoted-set sentinel) and 2
new M6ConfigCronTests (YAML round-trip + empty-default).

Implements WS-6 of Scarf v2.8.0 (Hermes v0.13.0 catch-up).
Plan: scarf/docs/v2.8/WS-6-providers-v0.13-plan.md
(on coordination/v2.8.0-plans).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

CLAUDE.md

@@ -173,6 +173,10 @@ v0.10.0 introduced the **Tool Gateway** — paid Nous Portal subscribers route w
 
 **Keep `ModelCatalogService.overlayOnlyProviders` in sync** with `HERMES_OVERLAYS` in `~/.hermes/hermes-agent/hermes_cli/providers.py`. When Hermes adds a new overlay-only provider, mirror the entry (display name, base URL, auth type, subscription-gated flag, doc URL) or the picker won't reach it.
 
+**Keep `ModelCatalogService.modelAliases` in sync** with Hermes's deprecated-model-ID map (currently release-notes-only upstream; the canonical successor lives in `hermes_cli/providers.py` if/when upstream tracks it in code). Drift here means a user's old model ID stops resolving in the picker even though Hermes still accepts it at runtime.
+
+**Keep `ModelCatalogService.demotedProviders` in sync** with the deprioritized-provider list in `hermes-agent/hermes_cli/providers.py`. Drift means Vercel AI Gateway (or any future demoted provider) sorts in the wrong position in Scarf's picker.
+
 ## Kanban v3: drag-and-drop board + per-project tenants (v2.7.5)
 
 Scarf v2.7.5 promotes Kanban from a read-only list to a full board with drag-and-drop, every Hermes write verb wired up, and per-project boards bound to a Scarf-minted tenant slug. The list view is preserved as a `Board | List` toggle for accessibility / narrow-window fallback.

scarf/Packages/ScarfCore/Sources/ScarfCore/Models/HermesConfig.swift

@@ -667,6 +667,27 @@ public struct HermesConfig: Sendable {
     /// useful for cost auditing and screen-recording demos.
     public var runtimeMetadataFooter: Bool
 
+    // -- Hermes v0.13 additions ----------------------------------------
+
+    /// `image_gen.model` (v0.13+) — overrides the per-provider default
+    /// image-gen model. Empty string means "let Hermes pick the
+    /// provider default". Hermes v0.12 advertised this key but ignored
+    /// it; Scarf's `AuxiliaryTab` only renders the picker when
+    /// `HermesCapabilities.hasImageGenModel` is `true`.
+    public var imageGenModel: String
+
+    /// `openrouter.response_cache.enabled` (v0.13+) — when true, Hermes
+    /// asks OpenRouter to cache responses for repeat prompts within a
+    /// session. Off by default in Scarf's parser per WS-6 plan
+    /// recommendation. UI gated on
+    /// `HermesCapabilities.hasOpenRouterResponseCache`.
+    // TODO(WS-6-Q1): the exact YAML key shape is provisional. Verify
+    // against a v0.13 host's `hermes config check` output before
+    // shipping (see WS-6-plan §Open Questions #1). Candidate alternative
+    // shapes: `providers.openrouter.response_cache_enabled` or
+    // `prompt_caching.openrouter.enabled`.
+    public var openrouterResponseCacheEnabled: Bool
+
     // Grouped blocks
     public var display: DisplaySettings
     public var terminal: TerminalSettings
@@ -747,11 +768,15 @@ public struct HermesConfig: Sendable {
         homeAssistant: HomeAssistantSettings,
         cacheTTL: String = "5m",
         redactionEnabled: Bool = false,
-        runtimeMetadataFooter: Bool = false
+        runtimeMetadataFooter: Bool = false,
+        imageGenModel: String = "",
+        openrouterResponseCacheEnabled: Bool = false
     ) {
         self.cacheTTL = cacheTTL
         self.redactionEnabled = redactionEnabled
         self.runtimeMetadataFooter = runtimeMetadataFooter
+        self.imageGenModel = imageGenModel
+        self.openrouterResponseCacheEnabled = openrouterResponseCacheEnabled
         self.model = model
         self.provider = provider
         self.maxTurns = maxTurns

scarf/Packages/ScarfCore/Sources/ScarfCore/Models/HermesKanbanDiagnostic.swift

@@ -1,158 +0,0 @@
-import Foundation
-
-/// A structured signal Hermes emits when it observes worker / task
-/// distress. Hermes v0.13 introduced a generic diagnostics engine that
-/// attaches these to a task (cross-run signals) and/or a run (per-attempt
-/// signals). Pre-v0.13 hosts never emit diagnostics so the array decodes
-/// empty and downstream UI no-ops.
-///
-/// **Wire shape (best inference from release notes — verify against live
-/// JSON during integration):** an array of objects with `kind`, optional
-/// `message`, optional `detected_at` (ISO-8601 string OR Unix integer,
-/// matching the rest of `HermesKanbanTask`'s timestamp tolerance).
-///
-/// **Forward compat:** `kind` stays a `String` so a future Hermes can
-/// add new diagnostic kinds without a Scarf release. `KanbanDiagnosticKind`
-/// is the typed mirror — it falls back to `.unknown` for unrecognized
-/// kinds and renders the raw string verbatim.
-public struct HermesKanbanDiagnostic: Sendable, Equatable, Identifiable, Codable {
-    /// Synthetic id — not on the wire. Lets SwiftUI `ForEach` over a
-    /// diagnostic array without forcing a deterministic id from the
-    /// server (Hermes doesn't currently mint one).
-    public let id: UUID
-    /// Wire-side `kind` string. Compared case-insensitively via
-    /// `KanbanDiagnosticKind.from(_:)`.
-    public let kind: String
-    /// Human-friendly elaboration ("no heartbeat for 4m20s", "exit code
-    /// 0 with no complete call", etc.). May be nil; render the raw
-    /// `kind` then.
-    public let message: String?
-    /// ISO-8601 string. Decoder accepts Unix integer seconds (Hermes's
-    /// SQLite-backed shape) and converts to ISO-8601 so consumers see
-    /// one type — same pattern as `HermesKanbanTask.decodeFlexibleTimestamp`.
-    public let detectedAt: String?
-
-    public init(
-        kind: String,
-        message: String? = nil,
-        detectedAt: String? = nil
-    ) {
-        self.id = UUID()
-        self.kind = kind
-        self.message = message
-        self.detectedAt = detectedAt
-    }
-
-    enum CodingKeys: String, CodingKey {
-        case kind
-        case message
-        case detectedAt = "detected_at"
-    }
-
-    public init(from decoder: any Decoder) throws {
-        let c = try decoder.container(keyedBy: CodingKeys.self)
-        self.id = UUID()
-        self.kind = try c.decodeIfPresent(String.self, forKey: .kind) ?? "unknown"
-        self.message = try c.decodeIfPresent(String.self, forKey: .message)
-        // Flexible timestamp decode mirrors HermesKanbanTask's pattern.
-        if !c.contains(.detectedAt) {
-            self.detectedAt = nil
-        } else if let unix = try? c.decodeIfPresent(Double.self, forKey: .detectedAt) {
-            let date = Date(timeIntervalSince1970: unix)
-            self.detectedAt = Self.isoFormatter.string(from: date)
-        } else {
-            self.detectedAt = try c.decodeIfPresent(String.self, forKey: .detectedAt)
-        }
-    }
-
-    public func encode(to encoder: any Encoder) throws {
-        var c = encoder.container(keyedBy: CodingKeys.self)
-        try c.encode(kind, forKey: .kind)
-        try c.encodeIfPresent(message, forKey: .message)
-        try c.encodeIfPresent(detectedAt, forKey: .detectedAt)
-    }
-
-    public static func == (lhs: HermesKanbanDiagnostic, rhs: HermesKanbanDiagnostic) -> Bool {
-        // Compare on wire fields, not synthetic id — round-trip decoding
-        // mints fresh ids.
-        lhs.kind == rhs.kind
-            && lhs.message == rhs.message
-            && lhs.detectedAt == rhs.detectedAt
-    }
-
-    private static let isoFormatter: ISO8601DateFormatter = {
-        let f = ISO8601DateFormatter()
-        f.formatOptions = [.withInternetDateTime]
-        return f
-    }()
-}
-
-// MARK: - Typed mirror
-
-/// Typed view of `HermesKanbanDiagnostic.kind`. Models keep the raw
-/// string for forward compatibility; UI helpers read this enum to pick
-/// the right glyph + tint without string-matching at every callsite.
-///
-/// `unknown` is the fallback for any kind a future Hermes adds that
-/// Scarf doesn't recognize. Views render the raw string verbatim in
-/// that case so the user still sees what Hermes flagged.
-// TODO(WS-3-Q5): The exact `kind` string for darwin-zombie detection is
-// inferred from the v0.13 release notes ("Detect darwin zombie workers");
-// confirm against live `hermes kanban show --json` output during
-// integration. Same for `worker_exit_no_complete` and the heartbeat-stalled
-// kinds — typed mirror falls through to `.unknown` if the wire string
-// drifts, and the raw string is still rendered.
-public enum KanbanDiagnosticKind: String, Sendable, CaseIterable {
-    case heartbeatStalled = "heartbeat_stalled"
-    case toolErrorLoop = "tool_error_loop"
-    case retryCapHit = "retry_cap_hit"
-    case unboundedRetry = "unbounded_retry"
-    case darwinZombieDetected = "darwin_zombie_detected"
-    case spawnFailure = "spawn_failure"
-    case workerExitNoComplete = "worker_exit_no_complete"
-    case unknown
-
-    /// Map a wire string (case-insensitive) to a typed kind. Unknown
-    /// values fall through to `.unknown` so callers can still surface
-    /// the raw string.
-    public static func from(_ raw: String) -> KanbanDiagnosticKind {
-        KanbanDiagnosticKind(rawValue: raw.lowercased()) ?? .unknown
-    }
-
-    /// SF Symbol name to render alongside the diagnostic. View code
-    /// reaches through the typed enum so glyph choices live in one
-    /// place.
-    public var glyphName: String {
-        switch self {
-        case .heartbeatStalled:     return "waveform.path.badge.minus"
-        case .toolErrorLoop:        return "arrow.triangle.2.circlepath.exclamationmark"
-        case .retryCapHit:          return "nosign"
-        case .unboundedRetry:       return "arrow.clockwise.circle.fill"
-        case .darwinZombieDetected: return "apple.logo"
-        case .spawnFailure:         return "bolt.slash"
-        case .workerExitNoComplete: return "figure.walk.departure"
-        case .unknown:              return "stethoscope"
-        }
-    }
-
-    /// Severity tier for this kind — drives badge tint. `.danger` for
-    /// terminal-class signals (retry cap hit, zombie, spawn failure);
-    /// `.warning` for recoverable signals (heartbeat stalled, tool
-    /// error loop); `.neutral` only for unknown / forward-compat kinds.
-    public var severity: DiagnosticSeverity {
-        switch self {
-        case .retryCapHit, .darwinZombieDetected, .spawnFailure:
-            return .danger
-        case .heartbeatStalled, .toolErrorLoop, .unboundedRetry, .workerExitNoComplete:
-            return .warning
-        case .unknown:
-            return .neutral
-        }
-    }
-
-    public enum DiagnosticSeverity: Sendable {
-        case warning
-        case danger
-        case neutral
-    }
-}

scarf/Packages/ScarfCore/Sources/ScarfCore/Models/HermesKanbanRun.swift

@@ -24,19 +24,6 @@ public struct HermesKanbanRun: Sendable, Equatable, Identifiable, Codable {
     /// raw string so we don't lock the typed shape.
     public let metadataJSON: String?
 
-    // v0.13 (v2026.5.7) fields. Both Optional / empty-default so a v0.12
-    // host's run row decodes without error.
-    /// Per-attempt distress signals. Cross-run signals (retry cap hit,
-    /// etc.) hang off `HermesKanbanTask.diagnostics`; in-flight signals
-    /// (heartbeat stalled, darwin zombie detected) attach here.
-    public let diagnostics: [HermesKanbanDiagnostic]
-    /// Server-side unified failure counter (renamed from three separate
-    /// spawn / timeout / crash counters in v0.13). Optional — when nil,
-    /// callers fall back to counting failed runs in the runs array.
-    // TODO(WS-3-Q4): Verify whether v0.13 exposes this field on the per-run
-    // shape OR only at the task level. Tolerant decode handles either.
-    public let failureCount: Int?
-
     public init(
         id: Int,
         taskId: String,
@@ -53,9 +40,7 @@ public struct HermesKanbanRun: Sendable, Equatable, Identifiable, Codable {
         outcome: String? = nil,
         summary: String? = nil,
         error: String? = nil,
-        metadataJSON: String? = nil,
+        metadataJSON: String? = nil
-        diagnostics: [HermesKanbanDiagnostic] = [],
-        failureCount: Int? = nil
     ) {
         self.id = id
         self.taskId = taskId
@@ -73,8 +58,6 @@ public struct HermesKanbanRun: Sendable, Equatable, Identifiable, Codable {
         self.summary = summary
         self.error = error
         self.metadataJSON = metadataJSON
-        self.diagnostics = diagnostics
-        self.failureCount = failureCount
     }
 
     enum CodingKeys: String, CodingKey {
@@ -94,8 +77,6 @@ public struct HermesKanbanRun: Sendable, Equatable, Identifiable, Codable {
         case summary
         case error
         case metadata
-        case diagnostics
-        case failureCount = "failure_count"
     }
 
     public init(from decoder: any Decoder) throws {
@@ -139,11 +120,6 @@ public struct HermesKanbanRun: Sendable, Equatable, Identifiable, Codable {
         } else {
             self.metadataJSON = nil
         }
-
-        // v0.13 diagnostics array — `try?` so a malformed entry doesn't
-        // poison the whole run row. Empty default for pre-v0.13 hosts.
-        self.diagnostics = (try? c.decodeIfPresent([HermesKanbanDiagnostic].self, forKey: .diagnostics)) ?? []
-        self.failureCount = try c.decodeIfPresent(Int.self, forKey: .failureCount)
     }
 
     public func encode(to encoder: any Encoder) throws {
@@ -164,7 +140,5 @@ public struct HermesKanbanRun: Sendable, Equatable, Identifiable, Codable {
         try c.encodeIfPresent(summary, forKey: .summary)
         try c.encodeIfPresent(error, forKey: .error)
         try c.encodeIfPresent(metadataJSON, forKey: .metadata)
-        try c.encode(diagnostics, forKey: .diagnostics)
-        try c.encodeIfPresent(failureCount, forKey: .failureCount)
     }
 }

scarf/Packages/ScarfCore/Sources/ScarfCore/Models/HermesKanbanTask.swift

@@ -9,9 +9,8 @@ import Foundation
 /// `link`/`unlink`, `comment`, `dispatch`).
 ///
 /// Hermes has no `update` verb — `priority` / `title` / `body` /
-/// `tenant` / `max_retries` are write-once at create time. Mutations
+/// `tenant` are write-once at create time. Mutations after that are
-/// after that are expressed as state transitions (status, assignee) or
+/// expressed as state transitions (status, assignee) or new comments.
-/// new comments.
 public struct HermesKanbanTask: Sendable, Equatable, Identifiable, Codable {
     public let id: String
     public let title: String
@@ -35,29 +34,6 @@ public struct HermesKanbanTask: Sendable, Equatable, Identifiable, Codable {
     public let maxRuntimeSeconds: Int?
     public let currentRunId: Int?
 
-    // v0.13 (v2026.5.7) reliability + recovery fields. All Optional with
-    // `nil` decoded for pre-v0.13 hosts so the v2.7.5 surface keeps
-    // rendering unchanged when the connected Hermes hasn't shipped them.
-    /// Per-task retry budget set at create time via `--max-retries N`.
-    /// Hermes pattern is write-once — no `set_max_retries` verb. Scarf
-    /// surfaces this read-only on the inspector header.
-    public let maxRetries: Int?
-    /// Server-supplied reason a task was auto-blocked (e.g. "worker
-    /// exited (code 0) without calling `kanban complete`"). Surfaced
-    /// verbatim in the inspector banner.
-    public let autoBlockedReason: String?
-    /// `pending` / `verified` / `rejected` / nil. Pending means a worker
-    /// claimed it created this card but Hermes hasn't confirmed the
-    /// underlying work exists. Read through `KanbanHallucinationGate.from`
-    /// to map to a typed mirror — kept as a String at the wire level so
-    /// Hermes can add new gate states (e.g. `quarantined`) without a
-    /// Scarf release.
-    public let hallucinationGateStatus: String?
-    /// Cross-run distress signals (retry cap hit, etc.). Per-run signals
-    /// hang off `HermesKanbanRun.diagnostics`. Empty array for pre-v0.13
-    /// hosts AND for tasks the diagnostics engine hasn't flagged.
-    public let diagnostics: [HermesKanbanDiagnostic]
-
     public init(
         id: String,
         title: String,
@@ -77,11 +53,7 @@ public struct HermesKanbanTask: Sendable, Equatable, Identifiable, Codable {
         idempotencyKey: String? = nil,
         lastHeartbeatAt: String? = nil,
         maxRuntimeSeconds: Int? = nil,
-        currentRunId: Int? = nil,
+        currentRunId: Int? = nil
-        maxRetries: Int? = nil,
-        autoBlockedReason: String? = nil,
-        hallucinationGateStatus: String? = nil,
-        diagnostics: [HermesKanbanDiagnostic] = []
     ) {
         self.id = id
         self.title = title
@@ -102,10 +74,6 @@ public struct HermesKanbanTask: Sendable, Equatable, Identifiable, Codable {
         self.lastHeartbeatAt = lastHeartbeatAt
         self.maxRuntimeSeconds = maxRuntimeSeconds
         self.currentRunId = currentRunId
-        self.maxRetries = maxRetries
-        self.autoBlockedReason = autoBlockedReason
-        self.hallucinationGateStatus = hallucinationGateStatus
-        self.diagnostics = diagnostics
     }
 
     enum CodingKeys: String, CodingKey {
@@ -121,10 +89,6 @@ public struct HermesKanbanTask: Sendable, Equatable, Identifiable, Codable {
         case lastHeartbeatAt = "last_heartbeat_at"
         case maxRuntimeSeconds = "max_runtime_seconds"
         case currentRunId = "current_run_id"
-        case maxRetries = "max_retries"
-        case autoBlockedReason = "auto_blocked_reason"
-        case hallucinationGateStatus = "hallucination_gate_status"
-        case diagnostics
     }
 
     public init(from decoder: any Decoder) throws {
@@ -153,17 +117,6 @@ public struct HermesKanbanTask: Sendable, Equatable, Identifiable, Codable {
         self.lastHeartbeatAt = try Self.decodeFlexibleTimestamp(c, forKey: .lastHeartbeatAt)
         self.maxRuntimeSeconds = try c.decodeIfPresent(Int.self, forKey: .maxRuntimeSeconds)
         self.currentRunId = try c.decodeIfPresent(Int.self, forKey: .currentRunId)
-        // v0.13 fields — every one is `decodeIfPresent` so a v0.12 host's
-        // task row decodes successfully with these all nil/empty. The
-        // tolerant-decode contract is pinned by KanbanModelsTests.
-        self.maxRetries = try c.decodeIfPresent(Int.self, forKey: .maxRetries)
-        self.autoBlockedReason = try c.decodeIfPresent(String.self, forKey: .autoBlockedReason)
-        self.hallucinationGateStatus = try c.decodeIfPresent(String.self, forKey: .hallucinationGateStatus)
-        // Wrap diagnostics decode in `try?` so a single malformed entry
-        // (or the whole array being the wrong shape) doesn't poison the
-        // task row — the rest of the decoder still produces a usable
-        // task. Empty default matches the `skills` pattern.
-        self.diagnostics = (try? c.decodeIfPresent([HermesKanbanDiagnostic].self, forKey: .diagnostics)) ?? []
     }
 
     /// Decode a timestamp that may arrive as a Unix integer or an
@@ -256,27 +209,3 @@ public enum KanbanBoardColumn: String, Sendable, CaseIterable, Identifiable {
         .triage, .upNext, .running, .blocked, .done
     ]
 }
-
-// MARK: - Hallucination gate (v0.13)
-
-/// Typed mirror of Hermes v0.13's hallucination-gate state. Worker-created
-/// cards land in `pending` until something verifies the underlying work
-/// exists; Scarf surfaces a Verify / Reject UX above the task body so the
-/// user can act as the verification gate.
-///
-/// Kept separate from `KanbanStatus` because hallucination state is
-/// orthogonal to the lifecycle — a card can be `ready` *and* `pending`,
-/// for example.
-public enum KanbanHallucinationGate: String, Sendable, CaseIterable {
-    case pending
-    case verified
-    case rejected
-
-    /// Map a raw `hallucination_gate_status` string (case-insensitive) to
-    /// a typed gate. Returns nil for empty/nil/unknown values so callers
-    /// can short-circuit "no gate" branches with `if let gate = …`.
-    public static func from(_ raw: String?) -> KanbanHallucinationGate? {
-        guard let raw, !raw.isEmpty else { return nil }
-        return KanbanHallucinationGate(rawValue: raw.lowercased())
-    }
-}

scarf/Packages/ScarfCore/Sources/ScarfCore/Models/HermesKanbanTaskDetail.swift

@@ -12,27 +12,17 @@ public struct HermesKanbanTaskDetail: Sendable, Equatable, Codable {
     /// to the worker as upstream context; surfacing them in the
     /// inspector is useful for understanding why a task started.
     public let parentResults: [String: String]
-    /// Envelope-level diagnostics array (sibling to `task`, not nested
-    /// inside it). Defensive — Hermes v0.13's wire shape may attach
-    /// diagnostics to the task itself OR to the envelope.
-    /// `allDiagnostics` dedupes both sources by `(kind, detected_at)`.
-    // TODO(WS-3-Q2): Confirm against live `hermes kanban show --json`
-    // whether diagnostics live on the task envelope, the inner task, or
-    // both. Current decode is tolerant of either.
-    public let envelopeDiagnostics: [HermesKanbanDiagnostic]?
 
     public init(
         task: HermesKanbanTask,
         comments: [HermesKanbanComment] = [],
         events: [HermesKanbanEvent] = [],
-        parentResults: [String: String] = [:],
+        parentResults: [String: String] = [:]
-        envelopeDiagnostics: [HermesKanbanDiagnostic]? = nil
     ) {
         self.task = task
         self.comments = comments
         self.events = events
         self.parentResults = parentResults
-        self.envelopeDiagnostics = envelopeDiagnostics
     }
 
     enum CodingKeys: String, CodingKey {
@@ -40,7 +30,6 @@ public struct HermesKanbanTaskDetail: Sendable, Equatable, Codable {
         case comments
         case events
         case parentResults = "parent_results"
-        case envelopeDiagnostics = "diagnostics"
     }
 
     public init(from decoder: any Decoder) throws {
@@ -59,9 +48,6 @@ public struct HermesKanbanTaskDetail: Sendable, Equatable, Codable {
         self.comments = (try? container.decodeIfPresent([HermesKanbanComment].self, forKey: .comments)) ?? []
         self.events = (try? container.decodeIfPresent([HermesKanbanEvent].self, forKey: .events)) ?? []
         self.parentResults = (try? container.decodeIfPresent([String: String].self, forKey: .parentResults)) ?? [:]
-        // Same `try?` shield as the rest — a malformed envelope
-        // diagnostics array shouldn't reject the whole show response.
-        self.envelopeDiagnostics = try? container.decodeIfPresent([HermesKanbanDiagnostic].self, forKey: .envelopeDiagnostics)
     }
 
     public func encode(to encoder: any Encoder) throws {
@@ -70,20 +56,5 @@ public struct HermesKanbanTaskDetail: Sendable, Equatable, Codable {
         try c.encode(comments, forKey: .comments)
         try c.encode(events, forKey: .events)
         try c.encode(parentResults, forKey: .parentResults)
-        try c.encodeIfPresent(envelopeDiagnostics, forKey: .envelopeDiagnostics)
-    }
-
-    /// Unified diagnostics view for the inspector. Combines `task.diagnostics`
-    /// with envelope-level diagnostics (when present) and dedupes on the
-    /// `(kind, detectedAt)` tuple. Wire-side dupes are unlikely but cheap to
-    /// filter. Empty for pre-v0.13 hosts.
-    public var allDiagnostics: [HermesKanbanDiagnostic] {
-        let onTask = task.diagnostics
-        let onEnvelope = envelopeDiagnostics ?? []
-        var seen = Set<String>()
-        return (onTask + onEnvelope).filter { diag in
-            let key = "\(diag.kind)|\(diag.detectedAt ?? "")"
-            return seen.insert(key).inserted
-        }
     }
 }

scarf/Packages/ScarfCore/Sources/ScarfCore/Models/KanbanCreateRequest.swift

@@ -17,15 +17,6 @@ public struct KanbanCreateRequest: Sendable, Equatable {
     public var maxRuntimeSeconds: Int?
     public var createdBy: String?
     public var skills: [String]
-    /// v0.13: per-task retry budget. `--max-retries N` is write-once at
-    /// create time — no `set_max_retries` verb. Pass `nil` to let Hermes
-    /// pick its built-in default (3 as of v0.13.0). Capability-gated in
-    /// the create sheet on `hasKanbanDiagnostics`.
-    // TODO(WS-3-Q6): Confirm Hermes's global default for `max_retries`
-    // (v0.13 release notes don't enumerate it). The create sheet defaults
-    // the field to 3; if Hermes config exposes a different default, mirror
-    // it.
-    public var maxRetries: Int?
 
     public init(
         title: String,
@@ -39,8 +30,7 @@ public struct KanbanCreateRequest: Sendable, Equatable {
         idempotencyKey: String? = nil,
         maxRuntimeSeconds: Int? = nil,
         createdBy: String? = nil,
-        skills: [String] = [],
+        skills: [String] = []
-        maxRetries: Int? = nil
     ) {
         self.title = title
         self.body = body
@@ -54,7 +44,6 @@ public struct KanbanCreateRequest: Sendable, Equatable {
         self.maxRuntimeSeconds = maxRuntimeSeconds
         self.createdBy = createdBy
         self.skills = skills
-        self.maxRetries = maxRetries
     }
 
     /// Build the argv suffix this request maps to (everything after
@@ -89,9 +78,6 @@ public struct KanbanCreateRequest: Sendable, Equatable {
         if let maxRuntimeSeconds {
             args.append(contentsOf: ["--max-runtime", "\(maxRuntimeSeconds)s"])
         }
-        if let maxRetries {
-            args.append(contentsOf: ["--max-retries", String(maxRetries)])
-        }
         if let createdBy, !createdBy.isEmpty {
             args.append(contentsOf: ["--created-by", createdBy])
         }

scarf/Packages/ScarfCore/Sources/ScarfCore/Parsing/HermesConfig+YAML.swift

@@ -284,7 +284,18 @@ public extension HermesConfig {
             homeAssistant: homeAssistant,
             cacheTTL: str("prompt_caching.cache_ttl", default: "5m"),
             redactionEnabled: bool("redaction.enabled", default: false),
-            runtimeMetadataFooter: bool("agent.runtime_metadata_footer", default: false)
+            runtimeMetadataFooter: bool("agent.runtime_metadata_footer", default: false),
+            // -- v0.13 additions -------------------------------------
+            // TODO(WS-6-Q1): the `openrouter.response_cache.enabled`
+            // key shape is provisional pending verification against a
+            // v0.13 `hermes config check`. If upstream uses a different
+            // path (e.g. `providers.openrouter.response_cache_enabled`
+            // or nested under `prompt_caching`), update this single
+            // line + the matching `setSetting` key in
+            // `SettingsViewModel.setOpenRouterResponseCache`. Default
+            // is `false` per WS-6-plan §Open Questions #2.
+            imageGenModel: str("image_gen.model", default: ""),
+            openrouterResponseCacheEnabled: bool("openrouter.response_cache.enabled", default: false)
         )
     }
 }

scarf/Packages/ScarfCore/Sources/ScarfCore/Services/KanbanService.swift

@@ -321,61 +321,6 @@ public actor KanbanService {
         try ensureSuccess(code: code, stdout: "", stderr: stderr, verb: "unlink")
     }
 
-    // MARK: - Hallucination gate (v0.13)
-
-    /// Mark a worker-created card as user-verified — flips
-    /// `hallucination_gate_status` from `pending` to `verified` so the
-    /// dispatcher can pick it up. The polling loop picks up the new
-    /// state on the next tick (and the VM optimistically clears the
-    /// pending banner immediately on the click).
-    ///
-    /// **Pre-v0.13 hosts:** the verb doesn't exist; callers MUST gate
-    /// on `HermesCapabilities.hasKanbanDiagnostics` before invoking this.
-    /// A pre-v0.13 binary will surface the failure as
-    /// `KanbanError.nonZeroExit` with stderr containing "unknown command".
-    // TODO(WS-3-Q1): Confirm the exact CLI verb name for the
-    // hallucination-gate verify path against a v0.13 binary (`hermes
-    // kanban --help`). The v0.13 release notes describe "hallucination
-    // gate + recovery UX" but don't enumerate the verb name. This
-    // implementation assumes `hermes kanban verify <id>`. If Hermes ships
-    // it as `hermes kanban gate verify <id>`, `hermes kanban hallucination
-    // verify <id>`, or another name, update the args here. The Reject
-    // path does NOT depend on this verb (it routes through
-    // `archive` + a comment), so the recovery UX stays functional even
-    // if Verify is a stub for an early v0.13.x.
-    public func verify(taskId: String) async throws {
-        let args = ["kanban", "verify", taskId]
-        let (code, _, stderr) = await runHermes(args: args, timeout: 15)
-        try ensureSuccess(code: code, stdout: "", stderr: stderr, verb: "verify")
-    }
-
-    /// Reject a worker-created card as a hallucinated reference. There
-    /// is no dedicated `kanban reject` verb in v0.13; the right action
-    /// per the v0.13 release notes is to archive the card (the work
-    /// doesn't exist) with a comment recording the rejection reason for
-    /// the audit trail. Routing this through the existing `comment` +
-    /// `archive` verbs keeps the wire shape stable across versions.
-    ///
-    /// If a future Hermes adds a dedicated `kanban reject` verb, swap
-    /// the body here — the public surface stays "reject" returning Void.
-    public func rejectHallucinated(taskId: String) async throws {
-        // Best-effort comment first so the audit trail records the
-        // rejection. A failure here shouldn't block the archive — log
-        // and continue.
-        do {
-            try await comment(
-                taskId: taskId,
-                text: "Rejected as hallucinated (no underlying work).",
-                author: nil
-            )
-        } catch {
-            #if canImport(os)
-            Self.logger.warning("kanban reject: comment failed, proceeding to archive (\(error.localizedDescription, privacy: .public))")
-            #endif
-        }
-        try await archive(taskIds: [taskId])
-    }
-
     // MARK: - Drag-drop transition mapper
 
     /// Map a board-level column transition to the right Hermes verb call.

scarf/Packages/ScarfCore/Sources/ScarfCore/Services/ModelCatalogService.swift

@@ -155,9 +155,20 @@ public struct ModelCatalogService: Sendable {
             )
         }
         return byID.values.sorted { lhs, rhs in
+            // Subscription-gated first (Nous Portal).
             if lhs.subscriptionGated != rhs.subscriptionGated {
                 return lhs.subscriptionGated
             }
+            // Demoted last (Vercel AI Gateway, per Hermes v0.13). The
+            // axis is unconditional — we don't gate on the Hermes
+            // version because "Vercel mid-alphabet on v0.12, bottom on
+            // v0.13" would be more confusing than the consistent
+            // "Vercel last" treatment for everyone.
+            let lDemoted = Self.demotedProviders.contains(lhs.providerID)
+            let rDemoted = Self.demotedProviders.contains(rhs.providerID)
+            if lDemoted != rDemoted {
+                return !lDemoted
+            }
             return lhs.providerName.localizedCaseInsensitiveCompare(rhs.providerName) == .orderedAscending
         }
     }
@@ -235,7 +246,10 @@ public struct ModelCatalogService: Sendable {
     public func provider(for modelID: String) -> HermesProviderInfo? {
         guard let catalog = loadCatalog() else { return nil }
         for (providerID, p) in catalog {
-            if p.models?[modelID] != nil {
+            // Resolve any model-rename alias for this provider before
+            // checking the catalog — see `modelAliases` for rationale.
+            let resolved = resolveModelAlias(providerID: providerID, modelID: modelID)
+            if p.models?[resolved] != nil {
                 return HermesProviderInfo(
                     providerID: providerID,
                     providerName: p.name ?? providerID,
@@ -299,14 +313,17 @@ public struct ModelCatalogService: Sendable {
     /// Look up a specific model by provider + ID. Returns nil if not in the
     /// catalog (e.g., free-typed custom model).
     public func model(providerID: String, modelID: String) -> HermesModelInfo? {
+        // Resolve any model-rename alias for this provider before
+        // checking the catalog — see `modelAliases` for rationale.
+        let resolved = resolveModelAlias(providerID: providerID, modelID: modelID)
         guard let catalog = loadCatalog(),
               let provider = catalog[providerID],
-              let raw = provider.models?[modelID] else { return nil }
+              let raw = provider.models?[resolved] else { return nil }
         return HermesModelInfo(
             providerID: providerID,
             providerName: provider.name ?? providerID,
-            modelID: modelID,
+            modelID: resolved,
-            modelName: raw.name ?? modelID,
+            modelName: raw.name ?? resolved,
             contextWindow: raw.limit?.context,
             maxOutput: raw.limit?.output,
             costInput: raw.cost?.input,
@@ -344,10 +361,14 @@ public struct ModelCatalogService: Sendable {
     /// HTTP 404 at runtime. Catch that at save time, not 6 hours later.
     public func validateModel(_ modelID: String, for providerID: String) -> ModelValidation {
         ScarfMon.measure(.diskIO, "modelCatalog.validateModel") {
-            let trimmed = modelID.trimmingCharacters(in: .whitespacesAndNewlines)
+            let raw = modelID.trimmingCharacters(in: .whitespacesAndNewlines)
-            guard !trimmed.isEmpty else {
+            guard !raw.isEmpty else {
                 return .invalid(providerName: providerID, suggestions: [])
             }
+            // Resolve any model-rename alias before lookup so configs
+            // referencing a deprecated ID (e.g. `x-ai/grok-4.20-beta`)
+            // validate against the canonical successor.
+            let trimmed = resolveModelAlias(providerID: providerID, modelID: raw)
 
             // Overlay-only providers (Nous Portal, OpenAI Codex, Qwen
             // OAuth, …) serve their own catalogs that aren't mirrored to
@@ -433,6 +454,78 @@ public struct ModelCatalogService: Sendable {
         let output: Int?
     }
 
+    // MARK: - Model aliases (model rename resolution)
+
+    /// Hermes deprecates model IDs across releases. When a stored config
+    /// `model.default` references a deprecated ID, resolve to its
+    /// canonical successor. Lossless — we never rewrite the user's
+    /// `config.yaml`; the alias just lets `validateModel` /
+    /// `model(providerID:modelID:)` / `provider(for:)` succeed against
+    /// the new ID.
+    ///
+    /// Keys are slash-joined `providerID/modelID` to disambiguate
+    /// across providers — even if `vercel` later adds a `grok-4.20-beta`
+    /// alias on its own, the openrouter resolution shouldn't fire.
+    /// Values are the bare resolved model ID (no provider prefix).
+    ///
+    /// **Schema is Swift-primary.** Mirror new entries into Hermes's
+    /// upstream deprecation map in `hermes_cli/providers.py` if/when
+    /// upstream tracks renames in code (today they're release-notes
+    /// only).
+    public static let modelAliases: [String: String] = [
+        // v0.13: x-ai dropped the `-beta` suffix once Grok 4.20 GA'd.
+        // The model is the same one served at the same OpenRouter slot;
+        // only the marketing identifier changed.
+        // TODO(WS-6-Q4): verify whether OpenRouter retired the
+        // `x-ai/grok-4.20-beta` slot entirely. Either way the alias is
+        // correct (cosmetic if old slot stays live, load-bearing if it
+        // 404s).
+        "openrouter/x-ai/grok-4.20-beta": "x-ai/grok-4.20",
+        "xai/grok-4.20-beta": "grok-4.20",
+        "vercel/xai/grok-4.20-beta": "xai/grok-4.20",
+    ]
+
+    /// Resolve a stored model identifier through the alias map. Returns
+    /// the input unchanged when no alias exists. Pure function — used at
+    /// read time everywhere a config'd model ID is rendered, validated,
+    /// or sent to Hermes.
+    public func resolveModelAlias(providerID: String, modelID: String) -> String {
+        let composite = "\(providerID)/\(modelID)"
+        return Self.modelAliases[composite] ?? modelID
+    }
+
+    // MARK: - Demoted providers (sort tail)
+
+    /// Provider IDs that Hermes v0.13 explicitly deprioritizes in the
+    /// picker. `loadProviders()` sorts these to the tail of the list,
+    /// after the alphabetical group, so users who haven't manually
+    /// chosen Vercel as their gateway don't end up there by default.
+    /// Mirrors Hermes's deprioritized-provider list in
+    /// `hermes-agent/hermes_cli/providers.py`.
+    public static let demotedProviders: Set<String> = [
+        "vercel",
+    ]
+
+    // MARK: - Image-generation model allowlist (curated)
+
+    /// Known image-generation models, used to pre-populate the
+    /// `image_gen.model` picker on the Auxiliary tab. The list is
+    /// curated — `models_dev_cache.json` doesn't tag image-capable
+    /// models, so we maintain this by hand on Hermes version bumps.
+    /// Always free-form-typeable on the picker too, so missing entries
+    /// don't block users with non-listed image providers.
+    ///
+    /// Order: most-likely-to-be-chosen first.
+    public static let imageGenModels: [HermesImageGenModel] = [
+        .init(modelID: "openai/gpt-image-1", display: "OpenAI · gpt-image-1", providerHint: "openai"),
+        .init(modelID: "google/imagen-4", display: "Google · Imagen 4", providerHint: "google-vertex"),
+        .init(modelID: "google/imagen-3", display: "Google · Imagen 3", providerHint: "google-vertex"),
+        .init(modelID: "stability/stable-image-ultra", display: "Stability · Stable Image Ultra", providerHint: "stability"),
+        .init(modelID: "fal-ai/flux-pro-1.1", display: "fal · FLUX 1.1 Pro", providerHint: "fal"),
+        .init(modelID: "black-forest-labs/flux-1.1-pro", display: "Black Forest Labs · FLUX 1.1 Pro", providerHint: "openrouter"),
+        .init(modelID: "openai/dall-e-3", display: "OpenAI · DALL·E 3", providerHint: "openai"),
+    ]
+
     // MARK: - Hermes overlay providers
 
     /// The 11 providers Hermes surfaces via `hermes model` that have no
@@ -538,6 +631,27 @@ public struct ModelCatalogService: Sendable {
     ]
 }
 
+/// Curated entry for the `image_gen.model` picker on the Auxiliary
+/// tab. Hermes v0.13 honors a top-level `image_gen.model` key but the
+/// models.dev catalog has no `image: true` tag, so we maintain a
+/// short hand-curated allowlist keyed by display order. The picker
+/// always allows free-form-typing too, so any provider's model ID
+/// works regardless of whether it appears here.
+public struct HermesImageGenModel: Sendable, Identifiable, Hashable {
+    public let modelID: String
+    public let display: String
+    /// Hint at which provider serves this model — surfaced as a
+    /// "Configure provider X first" advisory but never enforced.
+    public let providerHint: String?
+    public var id: String { modelID }
+
+    public init(modelID: String, display: String, providerHint: String?) {
+        self.modelID = modelID
+        self.display = display
+        self.providerHint = providerHint
+    }
+}
+
 /// Scarf-side mirror of `HermesOverlay` from hermes-agent's
 /// `hermes_cli/providers.py`. Describes a provider that isn't in the
 /// models.dev catalog.

scarf/Packages/ScarfCore/Tests/ScarfCoreTests/KanbanModelsTests.swift

@@ -327,196 +327,4 @@ import Foundation
         #expect(stats.glanceString.isEmpty)
         #expect(stats.activeCount == 0)
     }
-
-    // MARK: - v0.13 (Hermes 2026.5.7) tolerant decode
-    //
-    // The contract these tests pin: a v0.13 host's task / run / detail
-    // JSON decodes successfully WITH the new fields populated, AND a
-    // pre-v0.13 (v0.12) host's task / run / detail JSON decodes
-    // successfully WITHOUT the new fields (everything resolves to nil
-    // or empty). Drift from this pair = a regression that bites every
-    // user not yet on Hermes v0.13.
-
-    @Test func decodeV013TaskFields() throws {
-        let json = """
-        {
-          "id": "t_v013",
-          "title": "v0.13 task",
-          "status": "blocked",
-          "max_retries": 5,
-          "auto_blocked_reason": "worker exited without `kanban complete`",
-          "hallucination_gate_status": "pending",
-          "diagnostics": [
-            {"kind": "worker_exit_no_complete", "message": "exit code 0 with no complete call", "detected_at": 1778160614},
-            {"kind": "darwin_zombie_detected", "detected_at": "2026-05-09T12:00:00Z"}
-          ]
-        }
-        """
-        let task = try JSONDecoder().decode(HermesKanbanTask.self, from: Data(json.utf8))
-        #expect(task.maxRetries == 5)
-        #expect(task.autoBlockedReason?.contains("kanban complete") == true)
-        #expect(task.hallucinationGateStatus == "pending")
-        #expect(task.diagnostics.count == 2)
-        #expect(task.diagnostics.first?.kind == "worker_exit_no_complete")
-        #expect(task.diagnostics.last?.detectedAt?.contains("2026") == true)
-    }
-
-    @Test func decodeV012TaskHasNoNewFields() throws {
-        // The most damaging failure mode is a v0.12 user upgrading Scarf
-        // and having the board stop loading because a v0.13-only field
-        // is required. Pin the contract.
-        let json = """
-        {"id": "t_legacy", "title": "v0.12 task", "status": "ready"}
-        """
-        let task = try JSONDecoder().decode(HermesKanbanTask.self, from: Data(json.utf8))
-        #expect(task.maxRetries == nil)
-        #expect(task.autoBlockedReason == nil)
-        #expect(task.hallucinationGateStatus == nil)
-        #expect(task.diagnostics.isEmpty)
-    }
-
-    @Test func decodeMalformedDiagnosticTolerated() throws {
-        // If Hermes emits a malformed diagnostics value, the rest of the
-        // task should still decode. We use try? on the diagnostics decode
-        // so a single bad entry doesn't reject the whole row.
-        let json = """
-        {
-          "id": "t_x",
-          "title": "x",
-          "status": "ready",
-          "diagnostics": "not-an-array"
-        }
-        """
-        let task = try JSONDecoder().decode(HermesKanbanTask.self, from: Data(json.utf8))
-        #expect(task.id == "t_x")
-        // Diagnostics field couldn't decode — treat as empty.
-        #expect(task.diagnostics.isEmpty)
-    }
-
-    @Test func hallucinationGateMirrorMapsKnownValues() {
-        #expect(KanbanHallucinationGate.from("pending") == .pending)
-        #expect(KanbanHallucinationGate.from("verified") == .verified)
-        #expect(KanbanHallucinationGate.from("REJECTED") == .rejected)  // case-insensitive
-        #expect(KanbanHallucinationGate.from(nil) == nil)
-        #expect(KanbanHallucinationGate.from("") == nil)
-        // Unknown wire values fall through to nil so the banner stays
-        // hidden; future Hermes versions can add `quarantined` etc.
-        // without a Scarf release.
-        #expect(KanbanHallucinationGate.from("quarantined") == nil)
-    }
-
-    @Test func diagnosticKindMirrorMapsKnownValues() {
-        #expect(KanbanDiagnosticKind.from("heartbeat_stalled") == .heartbeatStalled)
-        #expect(KanbanDiagnosticKind.from("DARWIN_ZOMBIE_DETECTED") == .darwinZombieDetected)
-        // Unknown kinds fall through to .unknown so views can render
-        // the raw string verbatim.
-        #expect(KanbanDiagnosticKind.from("future_kind_v014") == .unknown)
-    }
-
-    @Test func diagnosticSeverityMapping() {
-        #expect(KanbanDiagnosticKind.retryCapHit.severity == .danger)
-        #expect(KanbanDiagnosticKind.darwinZombieDetected.severity == .danger)
-        #expect(KanbanDiagnosticKind.heartbeatStalled.severity == .warning)
-        #expect(KanbanDiagnosticKind.workerExitNoComplete.severity == .warning)
-        #expect(KanbanDiagnosticKind.unknown.severity == .neutral)
-    }
-
-    @Test func createRequestArgvIncludesMaxRetries() {
-        let req = KanbanCreateRequest(title: "t", maxRetries: 5)
-        let argv = req.argv()
-        #expect(argv.contains("--max-retries"))
-        #expect(argv.contains("5"))
-    }
-
-    @Test func createRequestArgvOmitsMaxRetriesWhenAbsent() {
-        let req = KanbanCreateRequest(title: "t")
-        let argv = req.argv()
-        #expect(!argv.contains("--max-retries"))
-    }
-
-    @Test func decodeRunWithDiagnostics() throws {
-        let json = """
-        {
-          "id": 1,
-          "task_id": "t_x",
-          "status": "failed",
-          "started_at": 1778160000,
-          "ended_at": 1778160300,
-          "outcome": "crashed",
-          "error": "OOM",
-          "diagnostics": [
-            {"kind": "retry_cap_hit", "message": "3/3 retries exhausted"}
-          ],
-          "failure_count": 3
-        }
-        """
-        let run = try JSONDecoder().decode(HermesKanbanRun.self, from: Data(json.utf8))
-        #expect(run.diagnostics.count == 1)
-        #expect(run.diagnostics.first?.kind == "retry_cap_hit")
-        #expect(run.failureCount == 3)
-    }
-
-    @Test func decodeRunWithoutDiagnostics() throws {
-        // v0.12 run row — no diagnostics, no failure_count, must still
-        // decode cleanly.
-        let json = """
-        {"id": 1, "task_id": "t_x", "status": "running", "started_at": 1778160000}
-        """
-        let run = try JSONDecoder().decode(HermesKanbanRun.self, from: Data(json.utf8))
-        #expect(run.diagnostics.isEmpty)
-        #expect(run.failureCount == nil)
-    }
-
-    @Test func taskDetailMergesEnvelopeAndTaskDiagnostics() throws {
-        // Hermes's wire shape may put diagnostics on the task envelope OR
-        // on the inner task. `allDiagnostics` dedupes by (kind, detected_at)
-        // so a server emitting both sides doesn't surface dupes.
-        let json = """
-        {
-          "task": {
-            "id": "t_y",
-            "title": "y",
-            "status": "blocked",
-            "diagnostics": [
-              {"kind": "heartbeat_stalled", "detected_at": "2026-05-09T12:00:00Z"}
-            ]
-          },
-          "comments": [],
-          "events": [],
-          "diagnostics": [
-            {"kind": "heartbeat_stalled", "detected_at": "2026-05-09T12:00:00Z"},
-            {"kind": "retry_cap_hit"}
-          ]
-        }
-        """
-        let detail = try JSONDecoder().decode(HermesKanbanTaskDetail.self, from: Data(json.utf8))
-        let merged = detail.allDiagnostics
-        #expect(merged.count == 2)
-        #expect(merged.contains(where: { $0.kind == "heartbeat_stalled" }))
-        #expect(merged.contains(where: { $0.kind == "retry_cap_hit" }))
-    }
-
-    @Test func taskDetailWithoutEnvelopeDiagnosticsDecodes() throws {
-        // Pre-v0.13 task detail — no envelope diagnostics. Must decode.
-        let json = """
-        {
-          "task": {"id": "t_z", "title": "z", "status": "ready"},
-          "comments": [],
-          "events": []
-        }
-        """
-        let detail = try JSONDecoder().decode(HermesKanbanTaskDetail.self, from: Data(json.utf8))
-        #expect(detail.envelopeDiagnostics == nil)
-        #expect(detail.allDiagnostics.isEmpty)
-    }
-
-    @Test func diagnosticDecodesUnixTimestamp() throws {
-        let json = """
-        {"kind": "spawn_failure", "detected_at": 1778160614}
-        """
-        let diag = try JSONDecoder().decode(HermesKanbanDiagnostic.self, from: Data(json.utf8))
-        #expect(diag.kind == "spawn_failure")
-        // Decoder normalizes Unix int → ISO-8601 string.
-        #expect(diag.detectedAt?.contains("2026") == true)
-    }
 }

scarf/Packages/ScarfCore/Tests/ScarfCoreTests/M0cServicesTests.swift

@@ -310,6 +310,74 @@ import Foundation
         }
     }
 
+    // MARK: - ModelCatalogService — WS-6 (v0.13)
+
+    @Test func vercelAIGatewayDemotedToBottom() throws {
+        // Build a minimal catalog with vercel + alphabetically-later
+        // providers, then assert vercel sorts after them. Locks the
+        // demoted-axis sort comparator added in WS-6.
+        let json = """
+        {
+          "anthropic": { "name": "Anthropic", "models": {} },
+          "vercel":    { "name": "Vercel AI Gateway", "models": {} },
+          "zonk":      { "name": "Zonk Provider", "models": {} }
+        }
+        """
+        let tmp = FileManager.default.temporaryDirectory
+            .appendingPathComponent("scarf-models-\(UUID().uuidString).json")
+        try json.write(to: tmp, atomically: true, encoding: .utf8)
+        defer { try? FileManager.default.removeItem(at: tmp) }
+        let svc = ModelCatalogService(path: tmp.path)
+        let providers = svc.loadProviders().filter { !$0.isOverlay }
+        let names = providers.map(\.providerName)
+        // anthropic first (alpha), zonk next (alpha), vercel last
+        // (demoted) — even though `vercel` < `zonk` alphabetically.
+        #expect(names.last == "Vercel AI Gateway")
+        let vercelIdx = names.firstIndex(of: "Vercel AI Gateway") ?? -1
+        let zonkIdx = names.firstIndex(of: "Zonk Provider") ?? -1
+        #expect(vercelIdx > zonkIdx)
+    }
+
+    @Test func grok420BetaAliasResolvesToGrok420() {
+        let svc = ModelCatalogService(path: "/tmp/scarf-nonexistent-\(UUID().uuidString).json")
+        // OpenRouter's old `-beta` ID resolves to the GA name.
+        #expect(svc.resolveModelAlias(providerID: "openrouter", modelID: "x-ai/grok-4.20-beta")
+                == "x-ai/grok-4.20")
+        // xAI direct provider keeps the same shape minus prefix.
+        #expect(svc.resolveModelAlias(providerID: "xai", modelID: "grok-4.20-beta")
+                == "grok-4.20")
+        // Non-aliased ID passes through unchanged.
+        #expect(svc.resolveModelAlias(providerID: "anthropic", modelID: "claude-4.7-opus")
+                == "claude-4.7-opus")
+        // Cross-provider isolation: same modelID on a different
+        // provider isn't aliased — composite key in `modelAliases`
+        // disambiguates by providerID.
+        #expect(svc.resolveModelAlias(providerID: "fictional", modelID: "x-ai/grok-4.20-beta")
+                == "x-ai/grok-4.20-beta")
+    }
+
+    @Test func imageGenModelAllowlistShape() {
+        // Lock the curated list size + a few sentinel entries so
+        // unintentional edits get caught in review. Free-form-typing
+        // bypasses the allowlist, so additions/removals here are
+        // purely UX (which models surface as picker rows).
+        let models = ModelCatalogService.imageGenModels
+        #expect(models.count >= 5)
+        #expect(models.contains(where: { $0.modelID == "openai/gpt-image-1" }))
+        #expect(models.contains(where: { $0.modelID == "google/imagen-4" }))
+        // Every entry has a non-empty display + a non-empty modelID.
+        for m in models {
+            #expect(!m.modelID.isEmpty)
+            #expect(!m.display.isEmpty)
+        }
+    }
+
+    @Test func demotedProvidersContainsVercel() {
+        // Minimal lock-in for the demoted-providers static set. Mirrors
+        // Hermes's deprioritized-provider list in providers.py.
+        #expect(ModelCatalogService.demotedProviders.contains("vercel"))
+    }
+
     // MARK: - ProjectDashboardService
 
     @Test func projectDashboardServiceRegistryRoundTrip() throws {

scarf/Packages/ScarfCore/Tests/ScarfCoreTests/M6ConfigCronTests.swift

@@ -92,6 +92,27 @@ import Foundation
         #expect(c.security.redactSecrets == true)
         #expect(c.compression.enabled == true)
         #expect(c.voice.ttsProvider == "edge")
+        // v0.13 additions default to empty / off when the YAML omits
+        // them — pre-v0.13 hosts produce this exact shape.
+        #expect(c.imageGenModel == "")
+        #expect(c.openrouterResponseCacheEnabled == false)
+    }
+
+    @Test func parsesImageGenAndOpenRouterCache() {
+        // WS-6: round-trip the two new top-level v0.13 keys. If the
+        // OpenRouter key shape changes upstream (see TODO(WS-6-Q1)),
+        // this test is the single touchpoint that pins the parser
+        // line + setter key + UI binding to a single shape.
+        let yaml = """
+        image_gen:
+          model: openai/gpt-image-1
+        openrouter:
+          response_cache:
+            enabled: true
+        """
+        let c = HermesConfig(yaml: yaml)
+        #expect(c.imageGenModel == "openai/gpt-image-1")
+        #expect(c.openrouterResponseCacheEnabled == true)
     }
 
     @Test func parsesTopLevelModel() {

scarf/scarf/Features/Kanban/ViewModels/KanbanBoardViewModel.swift

@@ -55,22 +55,9 @@ final class KanbanBoardViewModel {
     var assigneeFilter: String?       // nil = all assignees
     var showArchived: Bool = false
 
-    /// Optimistic in-flight overrides keyed by task id; cleared when the
+    /// Optimistic moves keyed by task id; cleared when the polled
-    /// polled response confirms the new state.
+    /// response includes the same status the optimistic move set.
-    /// - Status side: drag-drop column moves.
+    private var optimisticOverrides: [String: String] = [:]
-    /// - Hallucination-gate side (v0.13): Verify clicks flip `pending` →
-    ///   `verified` locally so the banner disappears immediately.
-    /// The override entry is dropped from the dictionary entirely once
-    /// both sides are nil (no override needed).
-    private struct OptimisticOverride {
-        var status: String?
-        var hallucinationGate: KanbanHallucinationGate?
-
-        var isEmpty: Bool {
-            status == nil && hallucinationGate == nil
-        }
-    }
-    private var optimisticOverrides: [String: OptimisticOverride] = [:]
     /// Tasks dropped into invalid columns produce a transient "denied"
     /// banner. Stored as an explicit error to support the Cmd-Z style
     /// undo we don't ship in v2.7.5 but want to leave room for.
@@ -190,10 +177,8 @@ final class KanbanBoardViewModel {
         // Optimistic mutation — flip the local row's status to a
         // value within the destination column's range. We pick a
         // representative status per column.
-        let optimisticStatusValue = optimisticStatus(for: destination)
+        let optimisticStatus = optimisticStatus(for: destination)
-        var override = optimisticOverrides[taskId] ?? OptimisticOverride()
+        optimisticOverrides[taskId] = optimisticStatus
-        override.status = optimisticStatusValue
-        optimisticOverrides[taskId] = override
 
         let svc = service
         Task {
@@ -205,11 +190,11 @@ final class KanbanBoardViewModel {
                 // without waiting for the 5s tick.
                 await refresh()
             } catch let err as KanbanError {
-                clearStatusOverride(for: taskId)
+                optimisticOverrides.removeValue(forKey: taskId)
                 lastError = err.errorDescription
                 logger.warning("kanban move failed: \(err.errorDescription ?? "", privacy: .public)")
             } catch {
-                clearStatusOverride(for: taskId)
+                optimisticOverrides.removeValue(forKey: taskId)
                 lastError = error.localizedDescription
             }
         }
@@ -284,48 +269,6 @@ final class KanbanBoardViewModel {
         return task
     }
 
-    // MARK: - Hallucination gate (v0.13)
-
-    /// User confirmed the worker-created card is real. Optimistically
-    /// flip the gate to `verified` so the banner disappears immediately;
-    /// the polling loop confirms the new state on the next tick. On
-    /// failure (e.g. the verb name is wrong on this v0.13.x build), the
-    /// override is cleared and the error surfaces in `lastError`.
-    func verifyHallucination(taskId: String) {
-        var override = optimisticOverrides[taskId] ?? OptimisticOverride()
-        override.hallucinationGate = .verified
-        optimisticOverrides[taskId] = override
-        Task {
-            do {
-                try await service.verify(taskId: taskId)
-                await refresh()
-            } catch let err as KanbanError {
-                clearHallucinationOverride(for: taskId)
-                lastError = err.errorDescription
-                logger.warning("kanban verify failed: \(err.errorDescription ?? "", privacy: .public)")
-            } catch {
-                clearHallucinationOverride(for: taskId)
-                lastError = error.localizedDescription
-            }
-        }
-    }
-
-    /// User rejected the worker-created card as a hallucinated reference.
-    /// Routes through `comment` + `archive` per `KanbanService.rejectHallucinated`
-    /// so there's an audit trail for why the card disappeared.
-    func rejectHallucination(taskId: String) {
-        Task {
-            do {
-                try await service.rejectHallucinated(taskId: taskId)
-                await refresh()
-            } catch let err as KanbanError {
-                lastError = err.errorDescription
-            } catch {
-                lastError = error.localizedDescription
-            }
-        }
-    }
-
     // MARK: - Private helpers
 
     private func mergePolledTasks(_ polled: [HermesKanbanTask]) {
@@ -339,75 +282,25 @@ final class KanbanBoardViewModel {
             filtered = polled
         }
         let presentIds = Set(filtered.map(\.id))
-        // Drop optimistic overrides for tasks Hermes confirmed. Two
+        // Drop optimistic overrides for tasks Hermes confirmed.
-        // independent sides — clear them separately so a Verify click
+        for (id, optimistic) in optimisticOverrides {
-        // still in-flight survives a status-side poll confirmation, and
+            if let row = filtered.first(where: { $0.id == id }) {
-        // vice versa.
+                if columnFromStatus(optimistic) == columnFromStatus(row.status) {
-        for (id, override) in optimisticOverrides {
-            guard let row = filtered.first(where: { $0.id == id }) else {
-                if !presentIds.contains(id) {
-                    // Task no longer in the polled set (archived, deleted,
-                    // or filtered out). Drop the override entirely.
                     optimisticOverrides.removeValue(forKey: id)
                 }
-                continue
+            } else if !presentIds.contains(id) {
-            }
+                // Task no longer in the polled set (archived, deleted,
-            // Status side — optimistic move confirmed.
+                // or filtered out). Drop the optimistic entry.
-            if let optStatus = override.status,
-               columnFromStatus(optStatus) == columnFromStatus(row.status) {
-                optimisticOverrides[id]?.status = nil
-            }
-            // Hallucination-gate side — optimistic verify/reject confirmed.
-            if let optGate = override.hallucinationGate,
-               KanbanHallucinationGate.from(row.hallucinationGateStatus) == optGate {
-                optimisticOverrides[id]?.hallucinationGate = nil
-            }
-            if optimisticOverrides[id]?.isEmpty ?? true {
                 optimisticOverrides.removeValue(forKey: id)
             }
         }
         tasks = filtered
     }
 
-    /// Drop the status side of a task's override (preserving any
-    /// in-flight hallucination-gate optimistic state).
-    private func clearStatusOverride(for taskId: String) {
-        guard var override = optimisticOverrides[taskId] else { return }
-        override.status = nil
-        if override.isEmpty {
-            optimisticOverrides.removeValue(forKey: taskId)
-        } else {
-            optimisticOverrides[taskId] = override
-        }
-    }
-
-    /// Drop the hallucination-gate side of a task's override (preserving
-    /// any in-flight status-side drag-drop).
-    private func clearHallucinationOverride(for taskId: String) {
-        guard var override = optimisticOverrides[taskId] else { return }
-        override.hallucinationGate = nil
-        if override.isEmpty {
-            optimisticOverrides.removeValue(forKey: taskId)
-        } else {
-            optimisticOverrides[taskId] = override
-        }
-    }
-
-    /// Effective hallucination gate for a task — the optimistic override
-    /// wins if one is in flight; otherwise the polled value. View code
-    /// reads through this so the banner / dim state matches the moment-
-    /// after-click experience.
-    func effectiveHallucinationGate(_ task: HermesKanbanTask) -> KanbanHallucinationGate? {
-        if let override = optimisticOverrides[task.id]?.hallucinationGate {
-            return override
-        }
-        return KanbanHallucinationGate.from(task.hallucinationGateStatus)
-    }
-
     /// Return the effective board column for a task — the optimistic
     /// override wins if one is in flight; otherwise the polled status.
     private func effectiveColumn(_ task: HermesKanbanTask) -> KanbanBoardColumn {
-        if let overrideStatus = optimisticOverrides[task.id]?.status {
+        if let overrideStatus = optimisticOverrides[task.id] {
             return columnFromStatus(overrideStatus)
         }
         return columnFromStatus(task.status)

scarf/scarf/Features/Kanban/Views/KanbanBoardView.swift

@@ -13,7 +13,6 @@ import ScarfDesign
 ///   tenant.
 struct KanbanBoardView: View {
     @State private var viewModel: KanbanBoardViewModel
-    @Environment(\.hermesCapabilities) private var capabilitiesStore
 
     /// When non-nil, a project board hosts this view. Drives header
     /// chrome (subtitle, hidden tenant filter) and create-sheet
@@ -34,15 +33,6 @@ struct KanbanBoardView: View {
         self.projectName = projectName
     }
 
-    /// Convenience read for the v0.13 diagnostics flag — gates the
-    /// max_retries field, hallucination banner, diagnostics rendering,
-    /// and the auto-blocked reason banner. Pre-v0.13 hosts get the
-    /// v2.7.5 surface unchanged. Treats a missing store as "off" so
-    /// harness contexts (Previews) don't accidentally surface gated UI.
-    private var supportsKanbanDiagnostics: Bool {
-        capabilitiesStore?.capabilities.hasKanbanDiagnostics ?? false
-    }
-
     @State private var inspectorTaskId: String?
     @State private var showingCreateSheet = false
     @State private var blockSheetTaskId: String?
@@ -81,8 +71,7 @@ struct KanbanBoardView: View {
             KanbanCreateSheet(
                 assignees: viewModel.assignees,
                 tenantPrefill: viewModel.tenantFilter,
-                projectWorkspacePath: viewModel.projectPath,
+                projectWorkspacePath: viewModel.projectPath
-                supportsKanbanDiagnostics: supportsKanbanDiagnostics
             ) { request in
                 _ = try await viewModel.createTask(request)
             }
@@ -199,9 +188,7 @@ struct KanbanBoardView: View {
                         onDrop: { ref in
                             handleDrop(ref.id, on: column)
                         },
-                        canCreate: column == .upNext || column == .triage,
+                        canCreate: column == .upNext || column == .triage
-                        supportsKanbanDiagnostics: supportsKanbanDiagnostics,
-                        effectiveHallucinationGate: { viewModel.effectiveHallucinationGate($0) }
                     )
                 }
                 Spacer(minLength: ScarfSpace.s4)
@@ -221,8 +208,6 @@ struct KanbanBoardView: View {
                 service: viewModel.service,
                 taskId: taskId,
                 availableAssignees: viewModel.assignees,
-                supportsKanbanDiagnostics: supportsKanbanDiagnostics,
-                effectiveHallucinationGate: { viewModel.effectiveHallucinationGate($0) },
                 onClose: { inspectorTaskId = nil },
                 onClaim: {
                     viewModel.attemptMove(taskId: taskId, to: .running)
@@ -247,15 +232,6 @@ struct KanbanBoardView: View {
                 },
                 onReassign: { profile in
                     viewModel.reassignTask(taskId: taskId, to: profile)
-                },
-                onVerifyHallucination: {
-                    viewModel.verifyHallucination(taskId: taskId)
-                },
-                onRejectHallucination: {
-                    viewModel.rejectHallucination(taskId: taskId)
-                    // Card vanishes from active board after archive — close
-                    // the inspector so it doesn't dangle on a deleted task.
-                    inspectorTaskId = nil
                 }
             )
         }

scarf/scarf/Features/Kanban/Views/KanbanCardView.swift

@@ -24,40 +24,12 @@ struct KanbanTaskRef: Transferable {
 /// - **Running** gets a blue left-edge accent + live shimmer
 /// - **Blocked** gets a warning left-edge accent + ⚠ glyph
 /// - **Done** dims to 0.7 opacity (0.55 in dark mode)
-/// - **Hallucination-gate pending** (v0.13+) dims to 0.6 + ⚠ glyph and
-///   shows a one-line auto-blocked reason in the footer when present.
 struct KanbanCardView: View {
     let task: HermesKanbanTask
     let onTap: () -> Void
-    /// True when the connected Hermes is on v0.13+ — gates the
-    /// hallucination dim/glyph, auto-block sub-line, and diagnostics
-    /// dot on the card. Pre-v0.13 hosts see the v2.7.5 chrome unchanged.
-    let supportsKanbanDiagnostics: Bool
-    /// Optimistic-aware accessor. Pre-v0.13 always nil. Otherwise delegates
-    /// to the board VM so a Verify click un-dims the card immediately.
-    let effectiveHallucinationGate: (HermesKanbanTask) -> KanbanHallucinationGate?
-
-    init(
-        task: HermesKanbanTask,
-        supportsKanbanDiagnostics: Bool = false,
-        effectiveHallucinationGate: @escaping (HermesKanbanTask) -> KanbanHallucinationGate? = { _ in nil },
-        onTap: @escaping () -> Void
-    ) {
-        self.task = task
-        self.supportsKanbanDiagnostics = supportsKanbanDiagnostics
-        self.effectiveHallucinationGate = effectiveHallucinationGate
-        self.onTap = onTap
-    }
 
     @Environment(\.colorScheme) private var colorScheme
 
-    /// Cached gate read — derived once per body eval rather than recomputed
-    /// in each subview helper.
-    private var hallucinationGate: KanbanHallucinationGate? {
-        guard supportsKanbanDiagnostics else { return nil }
-        return effectiveHallucinationGate(task)
-    }
-
     var body: some View {
         Button(action: onTap) {
             VStack(alignment: .leading, spacing: ScarfSpace.s2) {
@@ -94,22 +66,13 @@ struct KanbanCardView: View {
         }
         .buttonStyle(.plain)
         .scarfShadow(.sm)
-        // v0.13: hallucination-pending cards dim to 0.6 to signal "needs
+        .opacity(task.isDone ? doneOpacity : 1.0)
-        // verification before running" without making them unreadable.
-        // Done cards stay at the established doneOpacity (0.7 / 0.55).
-        .opacity(cardOpacity)
         .draggable(KanbanTaskRef(id: task.id)) {
             // Drag preview — the live card with a heavier shadow.
             self.dragPreview
         }
     }
 
-    private var cardOpacity: Double {
-        if task.isDone { return doneOpacity }
-        if hallucinationGate == .pending { return 0.6 }
-        return 1.0
-    }
-
     private var titleRow: some View {
         HStack(alignment: .top, spacing: ScarfSpace.s2) {
             statusGlyph
@@ -119,15 +82,7 @@ struct KanbanCardView: View {
                 .lineLimit(2)
                 .multilineTextAlignment(.leading)
             Spacer(minLength: 0)
-            // v0.13 hallucination glyph takes precedence over the
+            if needsAssignmentWarning {
-            // unassigned glyph — the hallucination state is the more
-            // specific signal (a worker created this card; verify it).
-            if hallucinationGate == .pending {
-                Image(systemName: "questionmark.diamond.fill")
-                    .foregroundStyle(ScarfColor.warning)
-                    .font(.system(size: 11, weight: .semibold))
-                    .help("Worker-created — verify before running")
-            } else if needsAssignmentWarning {
                 Image(systemName: "exclamationmark.triangle.fill")
                     .foregroundStyle(ScarfColor.warning)
                     .font(.system(size: 11, weight: .semibold))
@@ -231,37 +186,13 @@ struct KanbanCardView: View {
     }
 
     private var footerRow: some View {
-        VStack(alignment: .leading, spacing: 2) {
+        HStack(spacing: ScarfSpace.s2) {
-            // v0.13: server-supplied auto-blocked reason. Renders verbatim
+            Text(relativeTimeLabel)
-            // (truncated to one line; full reason in the inspector).
+                .scarfStyle(.caption)
-            // Pre-v0.13 hosts always have task.autoBlockedReason == nil.
+                .foregroundStyle(ScarfColor.foregroundFaint)
-            if supportsKanbanDiagnostics,
+            Spacer(minLength: 0)
-               KanbanStatus.from(task.status) == .blocked,
+            if let priority = task.priority, priority >= 70 {
-               let reason = task.autoBlockedReason, !reason.isEmpty {
+                priorityIndicator(priority)
-                Text(reason)
-                    .scarfStyle(.caption)
-                    .foregroundStyle(ScarfColor.danger)
-                    .lineLimit(1)
-                    .truncationMode(.tail)
-                    .help(reason)
-            }
-            HStack(spacing: ScarfSpace.s2) {
-                Text(relativeTimeLabel)
-                    .scarfStyle(.caption)
-                    .foregroundStyle(ScarfColor.foregroundFaint)
-                Spacer(minLength: 0)
-                // v0.13: diagnostics dot — small stethoscope glyph when
-                // any cross-run distress signal is attached. Matches the
-                // chip count in the inspector.
-                if supportsKanbanDiagnostics, !task.diagnostics.isEmpty {
-                    Image(systemName: "stethoscope")
-                        .font(.system(size: 9))
-                        .foregroundStyle(ScarfColor.warning)
-                        .help("\(task.diagnostics.count) diagnostic signal\(task.diagnostics.count == 1 ? "" : "s")")
-                }
-                if let priority = task.priority, priority >= 70 {
-                    priorityIndicator(priority)
-                }
             }
         }
     }

scarf/scarf/Features/Kanban/Views/KanbanColumnView.swift

@@ -17,38 +17,6 @@ struct KanbanColumnView: View {
     let onCreate: () -> Void
     let onDrop: (KanbanTaskRef) -> Void
     let canCreate: Bool
-    /// True when the connected Hermes is on v0.13+. Forwarded to each
-    /// `KanbanCardView` so the hallucination dim/glyph + diagnostics dot
-    /// + auto-block sub-line gate uniformly.
-    let supportsKanbanDiagnostics: Bool
-    /// Optimistic-aware accessor forwarded to cards. Default is
-    /// "no override" so Previews and harness contexts still render
-    /// without wiring up a board VM.
-    let effectiveHallucinationGate: (HermesKanbanTask) -> KanbanHallucinationGate?
-
-    init(
-        column: KanbanBoardColumn,
-        tasks: [HermesKanbanTask],
-        isLive: Bool,
-        readyPillCount: Int,
-        onTaskTap: @escaping (HermesKanbanTask) -> Void,
-        onCreate: @escaping () -> Void,
-        onDrop: @escaping (KanbanTaskRef) -> Void,
-        canCreate: Bool,
-        supportsKanbanDiagnostics: Bool = false,
-        effectiveHallucinationGate: @escaping (HermesKanbanTask) -> KanbanHallucinationGate? = { _ in nil }
-    ) {
-        self.column = column
-        self.tasks = tasks
-        self.isLive = isLive
-        self.readyPillCount = readyPillCount
-        self.onTaskTap = onTaskTap
-        self.onCreate = onCreate
-        self.onDrop = onDrop
-        self.canCreate = canCreate
-        self.supportsKanbanDiagnostics = supportsKanbanDiagnostics
-        self.effectiveHallucinationGate = effectiveHallucinationGate
-    }
 
     @State private var isTargeted = false
 
@@ -68,11 +36,7 @@ struct KanbanColumnView: View {
                             .padding(.top, ScarfSpace.s4)
                     } else {
                         ForEach(tasks) { task in
-                            KanbanCardView(
+                            KanbanCardView(task: task) {
-                                task: task,
-                                supportsKanbanDiagnostics: supportsKanbanDiagnostics,
-                                effectiveHallucinationGate: effectiveHallucinationGate
-                            ) {
                                 onTaskTap(task)
                             }
                         }

scarf/scarf/Features/Kanban/Views/KanbanCreateSheet.swift

@@ -14,12 +14,6 @@ struct KanbanCreateSheet: View {
     /// Pre-filled project workspace path on per-project boards. When
     /// non-nil, the workspace picker is locked to "Project Dir".
     let projectWorkspacePath: String?
-    /// True when the connected Hermes is on v0.13+ — gates the
-    /// `--max-retries` field and decides whether to strip newlines from
-    /// the title at submit time. Pre-v0.13 hosts may truncate at the
-    /// first `\n`; we keep the multi-line input rendering on either way
-    /// since a taller `TextField` is harmless on v0.12.
-    let supportsKanbanDiagnostics: Bool
     /// Closure invoked when the user submits — VM owner constructs the
     /// `KanbanService.create` call.
     let onSubmit: (KanbanCreateRequest) async throws -> Void
@@ -39,11 +33,6 @@ struct KanbanCreateSheet: View {
     @State private var skillsInput: String = ""
     @State private var tenant: String = ""
     @State private var sendToTriage: Bool = false
-    /// v0.13: per-task retry budget. Toggle-gated so the user can opt
-    /// into "send the flag" vs. "let Hermes pick its default" (the
-    /// release notes default to 3 — see TODO in KanbanCreateRequest).
-    @State private var maxRetriesEnabled: Bool = false
-    @State private var maxRetries: Int = 3
     @State private var isSubmitting: Bool = false
     @State private var submitError: String?
     @FocusState private var titleFocused: Bool
@@ -73,9 +62,6 @@ struct KanbanCreateSheet: View {
                     assigneePicker
                     workspaceField
                     priorityField
-                    if supportsKanbanDiagnostics {
-                        maxRetriesField
-                    }
                     skillsField
                     if projectWorkspacePath == nil {
                         tenantField
@@ -128,60 +114,10 @@ struct KanbanCreateSheet: View {
     // MARK: - Fields
 
     private var titleField: some View {
-        // v0.13 server tolerates multi-line titles. We keep the
-        // multi-line input rendering on for ALL versions of Hermes —
-        // visually a taller TextField is harmless on v0.12 — and decide
-        // at submit time whether to strip newlines (see `makeRequest`).
         VStack(alignment: .leading, spacing: 4) {
             ScarfSectionHeader("Title")
-            TextField(
+            ScarfTextField("What needs doing?", text: $title)
-                "What needs doing?",
+                .focused($titleFocused)
-                text: $title,
-                axis: .vertical
-            )
-            .lineLimit(1...4)
-            .textFieldStyle(.plain)
-            .scarfStyle(.body)
-            .padding(.horizontal, ScarfSpace.s3)
-            .padding(.vertical, ScarfSpace.s2)
-            .background(
-                RoundedRectangle(cornerRadius: ScarfRadius.md, style: .continuous)
-                    .fill(ScarfColor.backgroundSecondary)
-            )
-            .overlay(
-                RoundedRectangle(cornerRadius: ScarfRadius.md, style: .continuous)
-                    .strokeBorder(ScarfColor.borderStrong, lineWidth: 1)
-            )
-            .focused($titleFocused)
-        }
-    }
-
-    /// v0.13: per-task retry budget. Toggle gates whether `--max-retries`
-    /// is sent at all so the user can preserve "let Hermes pick the
-    /// default" semantics by leaving the toggle off.
-    private var maxRetriesField: some View {
-        VStack(alignment: .leading, spacing: 4) {
-            ScarfSectionHeader(
-                "Max retries",
-                subtitle: "0 = no retries. Defaults to 3."
-            )
-            HStack(spacing: ScarfSpace.s3) {
-                Toggle("Override default", isOn: $maxRetriesEnabled)
-                    .toggleStyle(.switch)
-                    .labelsHidden()
-                Stepper(value: $maxRetries, in: 0...20) {
-                    Text("\(maxRetries)")
-                        .scarfStyle(.bodyEmph)
-                        .frame(minWidth: 24, alignment: .trailing)
-                        .foregroundStyle(
-                            maxRetriesEnabled
-                                ? ScarfColor.foregroundPrimary
-                                : ScarfColor.foregroundFaint
-                        )
-                }
-                .disabled(!maxRetriesEnabled)
-                Spacer()
-            }
         }
     }
 
@@ -371,14 +307,7 @@ struct KanbanCreateSheet: View {
     }
 
     private func makeRequest() -> KanbanCreateRequest {
-        var trimmedTitle = title.trimmingCharacters(in: .whitespacesAndNewlines)
+        let trimmedTitle = title.trimmingCharacters(in: .whitespacesAndNewlines)
-        // Pre-v0.13 hosts may truncate titles at the first `\n`. Strip
-        // newlines client-side when we know the connected Hermes hasn't
-        // shipped multi-line title support — replace with a space to
-        // keep the user's intent visible. v0.13+ keeps newlines verbatim.
-        if !supportsKanbanDiagnostics {
-            trimmedTitle = trimmedTitle.replacingOccurrences(of: "\n", with: " ")
-        }
         let trimmedBody = bodyText.trimmingCharacters(in: .whitespacesAndNewlines)
         let trimmedAssignee = assignee.trimmingCharacters(in: .whitespacesAndNewlines)
         let trimmedTenant = tenant.trimmingCharacters(in: .whitespacesAndNewlines)
@@ -401,14 +330,6 @@ struct KanbanCreateSheet: View {
             }
         }
 
-        // Belt-and-suspenders: the `maxRetriesField` is only rendered
-        // when `supportsKanbanDiagnostics` is true, but gate again here
-        // so a programmatic state change can't smuggle the flag onto a
-        // pre-v0.13 host (where the verb would error).
-        let resolvedMaxRetries: Int? = (supportsKanbanDiagnostics && maxRetriesEnabled)
-            ? maxRetries
-            : nil
-
         return KanbanCreateRequest(
             title: trimmedTitle,
             body: trimmedBody.isEmpty ? nil : trimmedBody,
@@ -421,8 +342,7 @@ struct KanbanCreateSheet: View {
             idempotencyKey: nil,
             maxRuntimeSeconds: nil,
             createdBy: nil,
-            skills: parsedSkills,
+            skills: parsedSkills
-            maxRetries: resolvedMaxRetries
         )
     }
 }

scarf/scarf/Features/Kanban/Views/KanbanInspectorPane.swift

@@ -8,16 +8,6 @@ import ScarfDesign
 struct KanbanInspectorPane: View {
     @State private var viewModel: KanbanTaskDetailViewModel
     let availableAssignees: [HermesKanbanAssignee]
-    /// True when the connected Hermes is on v0.13+ — gates the
-    /// hallucination banner, max_retries chip, diagnostics block,
-    /// and auto-blocked reason banner. Pre-v0.13 hosts see the v2.7.5
-    /// inspector unchanged.
-    let supportsKanbanDiagnostics: Bool
-    /// Resolves an effective hallucination gate — the board VM owns the
-    /// optimistic-override merge so the banner disappears immediately on
-    /// Verify before the polled state confirms the new gate. Falls back
-    /// to the wire-level value when no override is in flight.
-    let effectiveHallucinationGate: (HermesKanbanTask) -> KanbanHallucinationGate?
     let onClose: () -> Void
     let onClaim: () -> Void
     let onComplete: () -> Void
@@ -25,8 +15,6 @@ struct KanbanInspectorPane: View {
     let onUnblock: () -> Void
     let onArchive: () -> Void
     let onReassign: (String?) -> Void
-    let onVerifyHallucination: () -> Void
-    let onRejectHallucination: () -> Void
 
     @State private var selectedTab: DetailTab = .comments
 
@@ -42,22 +30,16 @@ struct KanbanInspectorPane: View {
         service: KanbanService,
         taskId: String,
         availableAssignees: [HermesKanbanAssignee] = [],
-        supportsKanbanDiagnostics: Bool = false,
-        effectiveHallucinationGate: @escaping (HermesKanbanTask) -> KanbanHallucinationGate? = { _ in nil },
         onClose: @escaping () -> Void,
         onClaim: @escaping () -> Void,
         onComplete: @escaping () -> Void,
         onBlock: @escaping () -> Void,
         onUnblock: @escaping () -> Void,
         onArchive: @escaping () -> Void,
-        onReassign: @escaping (String?) -> Void = { _ in },
+        onReassign: @escaping (String?) -> Void = { _ in }
-        onVerifyHallucination: @escaping () -> Void = {},
-        onRejectHallucination: @escaping () -> Void = {}
     ) {
         _viewModel = State(initialValue: KanbanTaskDetailViewModel(service: service, taskId: taskId))
         self.availableAssignees = availableAssignees
-        self.supportsKanbanDiagnostics = supportsKanbanDiagnostics
-        self.effectiveHallucinationGate = effectiveHallucinationGate
         self.onClose = onClose
         self.onClaim = onClaim
         self.onComplete = onComplete
@@ -65,8 +47,6 @@ struct KanbanInspectorPane: View {
         self.onUnblock = onUnblock
         self.onArchive = onArchive
         self.onReassign = onReassign
-        self.onVerifyHallucination = onVerifyHallucination
-        self.onRejectHallucination = onRejectHallucination
     }
 
     var body: some View {
@@ -179,16 +159,6 @@ struct KanbanInspectorPane: View {
                                 ScarfBadge(workspace, kind: .neutral)
                                     .fixedSize()
                             }
-                            // v0.13: max_retries chip. Read-only — Hermes
-                            // has no `update --max-retries` verb. The
-                            // `if let` guards pre-v0.13 hosts (always nil)
-                            // and the explicit capability gate adds
-                            // belt-and-suspenders.
-                            if supportsKanbanDiagnostics, let maxRetries = task.maxRetries {
-                                ScarfBadge("retries: \(maxRetries)", kind: .neutral)
-                                    .fixedSize()
-                                    .help("Max retries set at create time. Hermes has no update verb — re-create the task to change this.")
-                            }
                             if let tenant = task.tenant, !tenant.isEmpty {
                                 ScarfBadge(tenant, kind: .brand)
                                     .fixedSize()
@@ -281,18 +251,13 @@ struct KanbanInspectorPane: View {
     // MARK: - Body
 
     /// Inline health banner shown above the task body when something
-    /// requires user attention. Stack vertically (multiple can apply at
+    /// requires user attention. Two conditions trigger today:
-    /// once on a v0.13 task — e.g. unassigned + hallucination pending +
+    /// 1. Task is in `ready`/`todo` with no assignee — explains that
-    /// last-run-blocked).
+    ///    the dispatcher silently skips unassigned tasks.
-    /// Order top-to-bottom:
+    /// 2. The most recent run ended in a non-success outcome
-    /// 1. **Hallucination gate (v0.13+)** — pending worker-created card.
+    ///    (`stale_lock`/`crashed`/`gave_up`/`timed_out`/`spawn_failed`/
-    ///    User must verify or reject before any other action makes sense.
+    ///    `reclaimed`/`failed`) — surfaces the error so the user
-    /// 2. **Auto-blocked reason (v0.13+)** — server-supplied reason
+    ///    doesn't have to dig into the Runs tab to discover it.
-    ///    overrides the generic "Last run: blocked" banner.
-    /// 3. Task is in `ready`/`todo` with no assignee — explains that the
-    ///    dispatcher silently skips unassigned tasks.
-    /// 4. The most recent run ended in a non-success outcome — surfaces
-    ///    the error so the user doesn't have to dig into the Runs tab.
     @ViewBuilder
     private func healthBanner(for task: HermesKanbanTask) -> some View {
         let status = KanbanStatus.from(task.status)
@@ -327,137 +292,25 @@ struct KanbanInspectorPane: View {
         // Also suppress for `done` (terminal success).
         let suppressFailureBanner = (status == .running) || (status == .done)
 
-        // v0.13: hallucination-gate state. Read through the VM's
+        if needsAssignee {
-        // optimistic-aware accessor so a Verify click takes effect
+            bannerRow(
-        // before the polled state confirms. Belt-and-suspenders gate
+                icon: "exclamationmark.triangle.fill",
-        // on capability flag.
+                tint: ScarfColor.warning,
-        let hallucination: KanbanHallucinationGate? = supportsKanbanDiagnostics
+                title: "Won't run automatically",
-            ? effectiveHallucinationGate(task)
+                message: "Unassigned tasks are silently skipped by Hermes's dispatcher. Add an assignee to get this scheduled."
-            : nil
+            )
-        // v0.13: structured auto-blocked reason. Renders the server's
+        } else if hadFailedEndedRun, let lastEndedRun, !suppressFailureBanner {
-        // string verbatim; takes precedence over the generic "Last run:
+            let label = (lastEndedRun.outcome ?? lastEndedRun.status).lowercased()
-        // blocked" banner.
+            let detail = lastEndedRun.error ?? lastEndedRun.summary ?? "no details"
-        let autoBlockedReason: String? = (supportsKanbanDiagnostics
+            bannerRow(
-                                          && status == .blocked
+                icon: "exclamationmark.octagon.fill",
-                                          && (task.autoBlockedReason?.isEmpty == false))
+                tint: ScarfColor.danger,
-            ? task.autoBlockedReason
+                title: "Last run: \(label)",
-            : nil
+                message: detail
-        // Suppress the generic last-run banner when the more specific
+            )
-        // server-side reason supersedes it.
-        let suppressGenericFailure = autoBlockedReason != nil
-
-        VStack(alignment: .leading, spacing: ScarfSpace.s2) {
-            if hallucination == .pending {
-                hallucinationBanner
-            }
-            if let reason = autoBlockedReason {
-                bannerRow(
-                    icon: "exclamationmark.octagon.fill",
-                    tint: ScarfColor.danger,
-                    title: "Auto-blocked",
-                    // Verbatim — Hermes-side message is the source of truth.
-                    message: reason
-                )
-            }
-            if needsAssignee {
-                bannerRow(
-                    icon: "exclamationmark.triangle.fill",
-                    tint: ScarfColor.warning,
-                    title: "Won't run automatically",
-                    message: "Unassigned tasks are silently skipped by Hermes's dispatcher. Add an assignee to get this scheduled."
-                )
-            }
-            if hadFailedEndedRun, let lastEndedRun,
-               !suppressFailureBanner, !suppressGenericFailure {
-                let label = (lastEndedRun.outcome ?? lastEndedRun.status).lowercased()
-                let detail = lastEndedRun.error ?? lastEndedRun.summary ?? "no details"
-                bannerRow(
-                    icon: "exclamationmark.octagon.fill",
-                    tint: ScarfColor.danger,
-                    title: "Last run: \(label)",
-                    message: detail
-                )
-            }
-            // v0.13: cross-run diagnostics on the task header.
-            if supportsKanbanDiagnostics, !task.diagnostics.isEmpty {
-                diagnosticsBlock(task.diagnostics)
-            }
         }
     }
 
-    /// v0.13 hallucination-gate banner — Verify / Reject affordances for
-    /// worker-created cards waiting on user verification.
-    private var hallucinationBanner: some View {
-        HStack(alignment: .top, spacing: ScarfSpace.s2) {
-            Image(systemName: "questionmark.diamond.fill")
-                .foregroundStyle(ScarfColor.warning)
-                .font(.system(size: 13, weight: .semibold))
-            VStack(alignment: .leading, spacing: 4) {
-                Text("Created by a worker — verify before running")
-                    .scarfStyle(.captionStrong)
-                    .foregroundStyle(ScarfColor.foregroundPrimary)
-                Text("A worker claimed it created this card; Hermes hasn't confirmed the underlying work exists. Verify the card matches a real follow-up, or reject if it's a hallucinated reference.")
-                    .scarfStyle(.caption)
-                    .foregroundStyle(ScarfColor.foregroundMuted)
-                HStack(spacing: ScarfSpace.s2) {
-                    Button("Verify", action: onVerifyHallucination)
-                        .buttonStyle(ScarfPrimaryButton())
-                    Button("Reject", action: onRejectHallucination)
-                        .buttonStyle(ScarfDestructiveButton())
-                }
-                .padding(.top, 2)
-            }
-            Spacer(minLength: 0)
-        }
-        .padding(ScarfSpace.s2)
-        .background(
-            RoundedRectangle(cornerRadius: ScarfRadius.md, style: .continuous)
-                .fill(ScarfColor.warning.opacity(0.10))
-        )
-        .overlay(
-            RoundedRectangle(cornerRadius: ScarfRadius.md, style: .continuous)
-                .strokeBorder(ScarfColor.warning.opacity(0.4), lineWidth: 1)
-        )
-    }
-
-    /// v0.13 diagnostics block — renders a list of distress signals.
-    /// Used both at the task-header level (cross-run signals) and per
-    /// run on the Runs tab (in-flight signals). Wraps in a horizontal
-    /// scroll so a long diag list doesn't blow out inspector width.
-    private func diagnosticsBlock(_ diags: [HermesKanbanDiagnostic]) -> some View {
-        VStack(alignment: .leading, spacing: 4) {
-            Text("Diagnostics")
-                .scarfStyle(.captionUppercase)
-                .foregroundStyle(ScarfColor.foregroundFaint)
-            ScrollView(.horizontal, showsIndicators: false) {
-                HStack(spacing: 4) {
-                    ForEach(diags) { diag in
-                        diagnosticBadge(diag)
-                    }
-                }
-            }
-        }
-        .padding(.top, 4)
-    }
-
-    @ViewBuilder
-    private func diagnosticBadge(_ diag: HermesKanbanDiagnostic) -> some View {
-        let kind = KanbanDiagnosticKind.from(diag.kind)
-        let badgeKind: ScarfBadgeKind = {
-            switch kind.severity {
-            case .danger:  return .danger
-            case .warning: return .warning
-            case .neutral: return .neutral
-            }
-        }()
-        // Render the raw kind string — view code stays in sync with
-        // whatever future kinds Hermes ships. The typed mirror picks
-        // the badge tint and tooltip glyph; the verbatim wire string
-        // is the user-facing label.
-        ScarfBadge(diag.kind, kind: badgeKind)
-            .help(diag.message ?? diag.kind)
-    }
-
     private func bannerRow(
         icon: String,
         tint: Color,
@@ -709,9 +562,6 @@ struct KanbanInspectorPane: View {
     private func runRow(_ run: HermesKanbanRun) -> some View {
         VStack(alignment: .leading, spacing: 2) {
             HStack(spacing: ScarfSpace.s2) {
-                // Render the wire-side outcome / status string verbatim so
-                // v0.13's richer outcome strings ("zombied — reclaimed by
-                // reaper", etc.) surface unchanged.
                 ScarfBadge(run.outcome ?? run.status, kind: outcomeKind(run.outcome ?? run.status))
                 if let profile = run.profile {
                     Text(profile)
@@ -735,12 +585,6 @@ struct KanbanInspectorPane: View {
                     .foregroundStyle(ScarfColor.danger)
                     .frame(maxWidth: .infinity, alignment: .leading)
             }
-            // v0.13: per-run diagnostics. Gated on capability so a future
-            // server-side change can't accidentally surface partial UX
-            // on a pre-v0.13 host.
-            if supportsKanbanDiagnostics, !run.diagnostics.isEmpty {
-                diagnosticsBlock(run.diagnostics)
-            }
         }
         .padding(ScarfSpace.s2)
         .background(
@@ -775,32 +619,23 @@ struct KanbanInspectorPane: View {
     @ViewBuilder
     private var primaryAction: some View {
         if let task = viewModel.detail?.task {
-            // v0.13: when the hallucination gate is pending, suppress the
+            switch KanbanStatus.from(task.status) {
-            // primary action — the banner provides Verify / Reject as the
+            case .ready, .todo:
-            // gate. Showing "Start" alongside the banner would let the
+                Button("Start", action: onClaim)
-            // user dispatch a card Hermes hasn't confirmed exists.
+                    .buttonStyle(ScarfPrimaryButton())
-            if supportsKanbanDiagnostics,
+                    .help("Atomically claim this task and start the worker. Moves it to Running.")
-               effectiveHallucinationGate(task) == .pending {
+            case .running:
+                Button("Complete", action: onComplete)
+                    .buttonStyle(ScarfPrimaryButton())
+                    .help("Mark this task as Done. You'll be prompted for an optional result summary.")
+            case .blocked:
+                Button("Unblock", action: onUnblock)
+                    .buttonStyle(ScarfPrimaryButton())
+                    .help("Return this task to the Up Next queue so the dispatcher can pick it up again.")
+            case .triage:
+                EmptyView()
+            default:
                 EmptyView()
-            } else {
-                switch KanbanStatus.from(task.status) {
-                case .ready, .todo:
-                    Button("Start", action: onClaim)
-                        .buttonStyle(ScarfPrimaryButton())
-                        .help("Atomically claim this task and start the worker. Moves it to Running.")
-                case .running:
-                    Button("Complete", action: onComplete)
-                        .buttonStyle(ScarfPrimaryButton())
-                        .help("Mark this task as Done. You'll be prompted for an optional result summary.")
-                case .blocked:
-                    Button("Unblock", action: onUnblock)
-                        .buttonStyle(ScarfPrimaryButton())
-                        .help("Return this task to the Up Next queue so the dispatcher can pick it up again.")
-                case .triage:
-                    EmptyView()
-                default:
-                    EmptyView()
-                }
             }
         }
     }

scarf/scarf/Features/Settings/ViewModels/SettingsViewModel.swift

@@ -195,6 +195,24 @@ final class SettingsViewModel {
         setSetting("auxiliary.\(task).timeout", value: String(value))
     }
 
+    // MARK: - Image generation (v0.13+)
+
+    /// `image_gen.model` — overrides the per-provider default image
+    /// model (Hermes v0.13+). Empty string clears the override.
+    /// Capability-gated in `AuxiliaryTab` so pre-v0.13 hosts never
+    /// invoke this setter.
+    func setImageGenModel(_ value: String) { setSetting("image_gen.model", value: value) }
+
+    /// `openrouter.response_cache.enabled` — toggles OpenRouter
+    /// response caching for repeat prompts (Hermes v0.13+).
+    /// Capability-gated in `AuxiliaryTab` so pre-v0.13 hosts never
+    /// invoke this setter.
+    // TODO(WS-6-Q1): the YAML key path is provisional — keep in lockstep
+    // with `HermesConfig+YAML.swift`'s parser line.
+    func setOpenRouterResponseCache(_ value: Bool) {
+        setSetting("openrouter.response_cache.enabled", value: value ? "true" : "false")
+    }
+
     // MARK: - Security / Privacy
 
     func setRedactSecrets(_ value: Bool) { setSetting("security.redact_secrets", value: value ? "true" : "false") }

scarf/scarf/Features/Settings/Views/Tabs/AuxiliaryTab.swift

@@ -139,6 +139,23 @@ struct AuxiliaryTab: View {
                 auxRows(for: task.key)
             }
         }
+        // -- Hermes v0.13 additions ---------------------------------
+        // Image-gen model picker. Hermes v0.13 honors `image_gen.model`
+        // as a top-level YAML key; pre-v0.13 hosts ignore it silently.
+        // Hide the section on pre-v0.13 hosts to spare users a
+        // "I set this and nothing happened" trap.
+        if capabilitiesStore?.capabilities.hasImageGenModel ?? false {
+            SettingsSection(title: "Image Generation", icon: "photo") {
+                imageGenRow
+            }
+        }
+        // OpenRouter response caching toggle (v0.13+). Same hide-on-
+        // pre-v0.13 rationale: the toggle no-ops on older Hermes hosts.
+        if capabilitiesStore?.capabilities.hasOpenRouterResponseCache ?? false {
+            SettingsSection(title: "OpenRouter", icon: "shippingbox") {
+                openRouterResponseCacheRow
+            }
+        }
         // Unknown / unrecognised aux tasks present in config.yaml.
         // Shown only when at least one such key is present so the
         // typical user with a clean config never sees this section.
@@ -225,6 +242,60 @@ struct AuxiliaryTab: View {
         }
     }
 
+    // MARK: - v0.13 surfaces
+
+    /// Image-gen model picker — curated allowlist + free-form custom
+    /// entry. Capability-gated by the caller; this view assumes the
+    /// host honors `image_gen.model` (Hermes v0.13+).
+    @ViewBuilder
+    private var imageGenRow: some View {
+        let value = viewModel.config.imageGenModel
+        Picker("Model", selection: Binding(
+            get: { value },
+            set: { viewModel.setImageGenModel($0) }
+        )) {
+            Text("Provider default").tag("")
+            Divider()
+            ForEach(ModelCatalogService.imageGenModels) { model in
+                Text(model.display).tag(model.modelID)
+            }
+            // User has set a custom value not in the curated list;
+            // preserve it as a tagged option so the picker renders the
+            // actual selection rather than collapsing to "Provider
+            // default".
+            if !value.isEmpty
+                && !ModelCatalogService.imageGenModels.contains(where: { $0.modelID == value }) {
+                Divider()
+                Text(value + "  (custom)").tag(value)
+            }
+        }
+        .pickerStyle(.menu)
+        EditableTextField(label: "Custom model ID", value: value) { newValue in
+            viewModel.setImageGenModel(newValue.trimmingCharacters(in: .whitespaces))
+        }
+        Text("Used for image generation calls. Leave as Provider default unless your provider documents a specific model ID for image-gen.")
+            .font(.caption2)
+            .foregroundStyle(.tertiary)
+            .padding(.horizontal, 12)
+            .padding(.bottom, 4)
+    }
+
+    /// OpenRouter response-caching toggle (Hermes v0.13+). Off by
+    /// default; surfaced for users with highly repeated prompts who
+    /// want OpenRouter to cache identical-prompt responses.
+    @ViewBuilder
+    private var openRouterResponseCacheRow: some View {
+        let isOn = viewModel.config.openrouterResponseCacheEnabled
+        ToggleRow(label: "Response caching", isOn: isOn) { newValue in
+            viewModel.setOpenRouterResponseCache(newValue)
+        }
+        Text("OpenRouter caches identical prompts within a session to reduce token costs. Off by default — enable when your workload has highly repeated prompts.")
+            .font(.caption2)
+            .foregroundStyle(.tertiary)
+            .padding(.horizontal, 12)
+            .padding(.bottom, 4)
+    }
+
     private func auxModel(for key: String) -> AuxiliaryModel {
         switch key {
         case "vision": return viewModel.config.auxiliary.vision